03-28-2012 10:51 AM
Say I have a data set with 4 character variables: x1-x4.
The data looks something like: ab cd ab ca.
I want to compare each variable to the other. Therefore, I have 6 combinations. I am interested in the ones that match (i.e., x1 matches x3, 'ab').
So my output should contain 3 columns:
First and second columns list the variables being compared to each other (total of 6 rows due to 6 combinations).
The third column stating whether or not there is a match.
I know to calculate the number of combinations, I use: c42=comb(4,2).
I am stuck thereafter. Any suggestions?
03-28-2012 11:48 AM
OK, I have added one more variable 'num' to identify the original obs no., take it out as you wish.
input (x1-x4) (:$);
aa ab aa bd
dd cc dd cc
data want (keep=first second match num) ;
array xx $ x1-x4;
do _i=1 to dim(xx);
do _j=_i+1 to dim(xx);
if xx(_i)=xx(_j) then match='Y';
03-28-2012 12:52 PM
Thank you very much! That helped a lot and it was very straightforward! I was able to tweak it and fit it to a more complicated data set.